Arifah Che Alhadi

A semantic retrieval of web documents using domain ontology

The Semantic Web vision offers the potential to express queries in a more semantic way. However, the unstructured nature of existing web documents, which lack semantics, proves to be a difficult task for such a query. To support this, the semantic information content of web documents needs to be specified in order to make the tangled information more structured and accessible. In this paper, we propose an approach meant to semantically query web documents using a natural language analysis technique and a domain specific ontology. Using both techniques, the tool gradually constructs the semantic document model of the documents retrieved from an existing search engine for each search session, which is represented as XML. The semantic model can then be semantically refined and browsed by the user.

Going beyond the surrounding text to semantically annotate and search digital images

Digital objects such as images and videos are fundamental resources in digital library. To assist in retrieving such objects usually they are being tagged by some keywords or sentences. The popular approach to tag digital objects is based on associated text. However, relying on associated text alone such as the surrounding text unable to semantically describe such objects. This paper discusses the use of WordNet and ConceptNet to tag digital images beyond terms available in the surrounding text. WordNet is used to disambiguate concepts or terms from the associated text and ConceptNet is meant to infer topics or common-sense knowledge from summarizing the text that describe the images. However, relying on WordNet alone is not sufficed particularly when it comes to disambiguate specific or domain dependent concepts. As such the Name Entity Recognition (NER) technique is required to annotate important entities such as name of a person, location and organization. Our work focused on on-lines news images that are richly described with textual description.

A collaborative names recommendation in the Twitter environment based on location

Friendster, Facebook, Twitter and many other microblogs have been introduced since 2004. These web 2.0 applications have become a powerful tool for communication. Each social web site has millions of users whose interact with each other regardless of their location and distance. Therefore, the mechanism of recommendation system for these sites is important for users to find suitable friends. Name recommendation should be made based on the concept of homophily which stated that relationships between individuals who have in common is higher than individuals who have nothing in common. Twitter is one of the popular social web sites that were developed in 2006. Many of the Twitter users are passive users. They just follow other users but on the other side they do not have many followers. This problem arises because reciprocal relationship is not required in Twitter. To overcome this problem, a recommendation system can help users in searching friends by taking into account reciprocal relationships. The main goal of this study is to use collaborative filtering techniques to recommend names based on geographical location. Users location is taken from the users profile by using coordinates of latitude and longitude. Celebrities profile data sets provided by the Korea Advanced Institute of Science and Technology (KAIST) are taken for testing purposes. The result of the testing indicates the potential of exploiting geographical locations in collaboratively recommending names within Twitter environment.

Towards Building Semantic Rich Model for Web Documents Using Domain Ontology

Accessing and extracting semantic meanings from web documents is crucial for the realization of Semantic Web. While the web offers the flexibility of making information easily available, it is considerably hard to find a fruitful way to describe, classify and present this information with rich semantic content. Therefore, the semantic information content of web documents need to be specified in order to make the tangled information more accessible to search engines and other applications. In this paper we propose an approach meant to assist in constructing semantic document models using natural language analysis technique and a domain specific ontology.

An Ensemble Similarity Model for Short Text Retrieval

The rapid growth of World Wide Web has extended Information Retrieval related technology such as queries for information needs become more easily accessible. One such platform is online question answering (QA). Online community can posting questions and get direct response for their special information needs using various platforms. It creates large unorganized repositories of valuable knowledge resources. Effective QA retrieval is required to make these repositories accessible to fulfill users information requests quickly. The repositories might contained similar questions and answer to users newly asked question. This paper explores the similarity-based models for the QA system to rank search result candidates. We used Damerau-Levenshtein distance and cosine similarity model to obtain ranking scores between the question posted by the registered user and a similar candidate questions in repository. Empirical experimental results indicate that our proposed ensemble models are very encouraging and give a significantly better similarity value to improve search ranking results.

A Domain Specific Ontology Driven to Semantic Document Modelling

To support the realisation of semantic web – as well as digital library, the semantic information content of web documents need to be specified in order to make the tangled information more accessible to search engines and other applications. A number of efforts to support the semantic representation of web documents have been proposed. One such effort is the semantic document modelling of which existing web documents are classified and organised to form a semantic document model representing the contents of respective web documents. In this paper we propose a tool meant to assist in constructing semantic document models using natural language analysis technique and a domain specific ontology. Together with users involvement and participation the tool gradually construct the semantic document model which is represented as XML.